Intelligence agencies can hurt the innocent. Intelligence workers should have a duty of care to everyone with whom they come into contact, not just to the human assets (spies) they recruit. This goes way beyond respecting the public’s “right to privacy”. How much this can be achieved is debatable.
February 2026. This is a revision of a 2020 article, triggered by a need to support AI developments and a deteriorating global security order. It excludes legal issues, which is a large and separate area.
Trust must be earned, not presumed
Doctors know our bodies and our bad life styles – we tell them because we trust them.
Commercial companies target adverts at us based on a wealth of information – we accept that in exchange for free services, such as web browsing. There is a level of transparency because we can see anomalies, such as when you’ve been near smart devices, talked about something new, and the next time you go online you find receive for it.
However, with intelligence agencies we don’t choose – it’s taken, we don’t know what they take, or why, or how they could use it against us.
It has improved in the past.
Laws have been introduced in many countries to limit the surveillance of their citizens, and there is some enforcement.It can improve again.
The state could publish indicators of the extent of surveillance, and give examples of how it helps them. Independent oversight bodies could assess the processes for its collection and use.
Friends and AI without scruples
Intelligence agencies trade in shared intelligence. But is this used to avoid laws and break norms (rules)?
- For example, staff in the British GCHQ must not intercept local correspondence for its own people, except when an intrusive surveillance order has been signed. But what stops American NSA monitoring Brits intrusively, and using it to warn of persons-of-interest?
- Likewise, agencies don’t need a court order to use AI to monitor us, if the raw intel is kept hidden. The AI judges potential suspects, then humans check it using permitted sources (such Open Source Intelligence), and if there’s sufficient risk, the case is escalated for intrusive surveillance. Upside: AI provides massive cost savings while expanding the size of the “wolf lists”. Downside: the AI will identify innocents as potential risks.
For an example story on illegal phone intercepts, see my 2-page story “Judicial oversight“.
The secret judgement of the innocent
Intrusive surveillance is not invisible surveillance. If you’re observant, you see strangers who don’t fit in, things that change on your computer, and perhaps occasional visits of a poltergeist to your home. If you’re considered a risk, it could continue for a long period.
- You’re innocent? Maybe, but you could come under surveillance because of someone you know, or somewhere you were, or something innocent you did that looked suspicious to the AI. The intelligence teams will only have snapshots on which to base their judgements about you. You will not be consulted before they judge you, their judgement may be biased, especially if it’s based on AI. You can’t contact them, or protest innocence.
- You’re prone to paranoia? Your suspicions will be magnified if you detect you’re under surveillance by the Deep State. It can feel deeply wrong.
- You’re guilty? You’ve now been warned.
See my 2-page story on intrusive surveillance authorisation: “Guilty“.
Do the opinions of secret intelligence workers matter if you are innocent? Potentially, yes. In the UK, they pass intelligence findings to their police contacts then disappear into the shadows. In the USA, the FBI handle both internal intelligence and policing roles. With either approach, you may eventually be seen to be innocent, but the journey could be traumatic.
If you’re under intrusive surveillance, and you know it, you may not know why. Most people have a tendency to guess at the worst. And given stories of rogue groups and black ops, it adds to the fear.
It has improved in the past.
Information exists about real practices, for internal surveillance. For example, two previous directors of the UK’s GCHQ have published books documenting intelligence practices. (Caveat: There’s lots of detail left out.)It can improve again.
This time, it needs support from people outside the intelligence community. Academics and authoritative writers need to be accurate. At least some fiction writers and dramatists should seek accuracy. (I try. See About Adrian – spy thriller author.) And the intelligence services should use their networks of trusted contacts to help smooth the path.
Related reading on “duty of care”
“Covert Human Intelligence Sources”. UK Secret Service (MI5) website. (Retrieved 02-Nov-2020.)
“Secret Agents and Covert Humans Sources”. Chapter 4 of Principled Spying, The Ethics of Secret Intelligence, by David Omand and Mark Phythian. Oxford University Press, 2018.